[Milsurplus] SRR-11/12/13 and subminis

[antqradio at juno.com]

Peter
I think it is not really a fair question.  It's comparing apples to
oranges.

Modern semiconductors have a high reliability in part because not much is
done to it mechanically.  It is still the "same" semiconductor substrate
after processing as it was before.  Except for the implantation process
where donor and (or) acceptor impurities are forced into the substrate,
the rest is etching and deposition to form the PN junctions, passive
components and interconnects.  Since there can be hundreds to tens of
thousands of devices on each wafer, all formed at the same time, chances
are good that if one is good, more then 90% of the rest will yield to
spec.  There are always those pesky half-circuits around the
circumference of the wafer!  The majority of semiconductors, including
ICs, are processed with less then 15 mask levels which are used to define
all of the circuit elements.

Not so with tubes, Nuvistors excepted.  There are about 300 different
materials used and many mechanical operations needed to form micas,
supports, cathodes, heaters, plates grids etc, all done sequentially. 
Lots of special materials and operations to mess up to degrade
reliability.

That said, there are some instances where tube life is measured in tens
of years of continuous operation.  Submarine telephone cables, space
probes and communications satellites are three examples that come to
mind.  Voyager's TWT is still operating thirty some years after launch,
the submarine telephone cables were abandoned for better technology but
they were still operating when shut down.  One repeater every ten to
thirty miles using three vacuum tubes each, IIRC.  These were very
expensive tubes (the cost of reliability)!

I checked out my copy of "Subminiature Electron Tube Life Factors" by
Edwards, Lammers and Zoellner of GE, published in 1961.  Quoting from
Chapter 8, Conclusions:

"In one such test involving 60 dual-section tubes, no catastrophic
failures had occurred and all characteristic were within initial limits
on every section after 16,000 hours of operation.  Also no significant
slumping of characteristics occurred in the last 5,000 hours of the test,
making it impossible to use extrapolation methods to estimate the end of
life due to characteristic degradation."

Interesting to note that the recommendations for long term characteristic
stability is to operate the heater at 5% to 12% below rating, optimum
bulb temperature of 150 degree C and plate dissipation at 10% to 70% of
rating.  Heater-Cathode potential not to exceed 50 volts, plate voltage
not to exceed 125 volts.
There was more.

Regards from Arkansas,
Jim


On Wed, 04 Jan 2006 08:19:20 -0500 Peter Gottlieb <nerd at verizon.net>
writes:
> Given the most modern manufacturing processes and materials, would it 
> be 
> possible to build tubes which had a longer lifetime than solid state 
> devices?
> 
> 40,000 hours is under 5 years of continuous service, nothing special 
> by today's 
> standards.
> 
> 
> Kenneth G. Gordon wrote:
> > 
> > Yes, but they take one hell of a beating in that few minutes and 
> have to 
> > be 100% reliable doing it.
> > 
> ______________________________________________________________
> Milsurplus mailing list
> Home: http://mailman.qth.net/mailman/listinfo/milsurplus
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Milsurplus at mailman.qth.net
> 
>